Incorrect hinge anchor in SimulatedPlatform.cpp

David L Jung — Wed, 17 Nov 2004 19:27:51 -0000

Andreas Gunther found a bug in the setAnchor call that
sets the hinge anchor vector for the non-holonomic
platform back wheels in SimulatedPlatform.cpp
It will cause strange behaviour of the second vehicle
in the sim_tut1 & 2 programs.
It has been fixed as below in the repository. He writes:

"It looks like the rotation axis is not anchored to the
center of the wheels. Looking at the
SimulatedPlatform::construct() function I found these
lines:
...
hingeLeft->setAnchor( leftBackWheelPos -
platformSolids[Platform]->getPosition() );
...
hingeRight->setAnchor( rightBackWheelPos -
platformSolids[Platform]->getPosition() );
...

There is a comment in the documentation of the
HingeJoint::setAnchor() function which says that the
coordinates of this point must be given in the
coordinate system of the first body attached to this
joint. But the result of the vector subtraction depends
on the orientation of the robot in global space which
should not be the case. Here is what I did to correct
the behaviour:
...
hingeLeft->setAnchor(
platformSolids[Platform]->getOrienation().invert().rotate(leftBackWheelPos
- platformSolids[Platform]->getPosition() ) );
...
"

AnalyticLagrangianFSBetaOptimizer nullSpace? args reversed

David L Jung — Thu, 10 Apr 2003 14:47:38 -0000

AnalyticLagrangianFSBetaOptimizer.cpp contains some
expressions which read nullSpace?1:0, but they should
read nullSpace?0:1.

Matrix classes operator+= & -= broken

David L Jung — Thu, 07 Feb 2002 06:17:48 -0000

This has been fixed in my development tree. Just be
aware of it.
It is fixed for the next snapshot (and the Matrix &
Vector classes have been completely rewritten).

Recent changes to bugs

Incorrect hinge anchor in SimulatedPlatform.cpp

AnalyticLagrangianFSBetaOptimizer nullSpace? args reversed

Matrix classes operator+= &amp; -= broken

Matrix classes operator+= & -= broken